130
Chemistry 1983
processes in an elegant series of studies involving the reactions of Pt(IV)
complexes with those of Pt( II). I n an important departure, Anet (54) showed
that the “capture” property of Cr
2+
(aq) in being oxidized can be exploited to
produce complexes in which an organic radical is ligated to Cr(III). An entire
chapter of the volume of reference 1 is devoted to the chemistry of similar
organochromium complexes(55).
ELECTRONIC STRUCTURE AND MECHANISM
A major theme in my own research, on returning from leave, was to try to
understand the large differences in rate, noted qualitatively in my early work
(29), and later made more quantitative, for the reactions Cr
2+
(aq) with carbox-
ylate complexes of (NH
3
)
5
Co(III). Many of the ligands were dicarboxylic
acids, and to explain the observation that when a conjugated bond system
connects the two carboxylates, reaction is usually more rapid than it is for the
saturated analog, it seemed reasonable then to assume that in the case of
conjugated ligands, the reducing agent attacks the exo carboxyl (remote at-
tack), the conjugated bond system serving as a “conduit” for electron transfer.
In adopting this view the tacit assumption was made that the reactions are
non-adiabatic so that the extent of electronic coupling would be reflected in the
rate. In retrospect, this assumption is naive, because the effect of conjugation
would be exerted even if the reducing agent attacked at the endo carboxyl.
A false start was made in demonstrating remote attack defined as above.
Activation effects accompanying electron transfer were reported (56), which if
true, would have constituted proof of remote attack for these systems. These
effects could not be reproduced (57) in later work (I had by now moved from
University of Chicago to Stanford University). Remote attack for the large
organic ligands was finally demonstrated (58) in the reaction of
a measurement of the rate of reaction of Cr
2+
(aq) with the analogous Cr(III)
species, provided the clue to understanding, at least in a qualitative way, the
rate differences observed for different conjugated ligands. The astonishing
result was that the rate of reduction of the Co(III) complex is only about 10-
fold greater than that of the Cr(III) complex. When the bridging group is a
non-reducible species such as acetate, the ratio is > 10
4
. The insensitivity of
rate to the nature of the oxidant suggested that the electron does not transfer
directly from Cr(II) to the oxidizing center but that the mechanism rather
involves the le
-
reduction (59) of the ligand by the strong reducing agent Cr
2+
,
followed by reduction of the oxidizing center by the organic radical - i.e., a
“hopping” mechanism obtains. This view provided satisfactory rationaliza-
tions of most of the observations of rates made with organic ligands. For
example, the fact that the rate is considerably greater for HO
2
C-CH=CH-
C O
2
-
(fumarate) than for CH
3
C O
2
- as ligand on Co(III), may have little to do
with the opportunity that Cr(II) has to attack the remote carboxyl in the
H. Taube
131
former case, and only reflects the fact that fumarate can be reduced by
Cr
2+
(aq). Moreover, the otherwise puzzling observation that the rate for the
fumarate complex is increased (60) by H
+
is now easily understood; positive
charge added to the ligand makes it easier to reduce. That many reactions of
the class under consideration proceed by a stepwise mechanism has been
convincingly demonstrated and amply illustrated in subsequent work, most of
it done by Gould and co-workers (61).
The rationalization offered for the operation of the stepwise mechanism in
the Co(III)-Cr(II) systems is that the carrier orbital on the ligand has
π
symmetry, while the donor and acceptor orbitals have
σ symmetry. This
renders as highly improbable an event in which the four conditions: Franck-
Condon restrictions at each center, and the symmetry restrictions at each
center, are simultaneously met. Whether or not this is the correct explanation,
it led me to search for an oxidizing center of the ammine class in which the
acceptor orbital has
π symmetry.
When the important condition that the complexes undergo substitution
slowly was added, only one couple within the entire periodic table, Ru
3+/2+
then qualified (62). In principle, the Os
3 + / 2 +
couple is also a
candidate, but unless some strong
π acid ligands are present, the couple is too
strongly reducing to be useful. The ruthenium species had the added advantage
that much more in the way of preparative work was known (63), and further
that the redox potentials are close to those of the much studied cobalt couples.
Since the
π orbital on Ru(III) can overlap with the π∗ orbital on the ligand, we
expected that the “hopping” mechanism would no longer obtain. Reaction
with Cr
2+
(aq) is in fact much more rapid (2 x 10
4
) than it is in the case of the
Co(III) isonicotinamide complex, and moreover, the rates are now quite
sensitive to changes in the redox potential of the oxidizing center (64). The
chemistry also differs in an interesting way from the Co(III)-Cr(II) case. The
bond between Ru and the ligand is not severed when Ru(III) is reduced to
Ru(II), and a kinetically stable binuclear intermediate is formed, as is expected
(18) from the electronic structures of the products,
for Ru(II) and
for
Cr(III).
Though the main intent of this paper is to provide historical background
rather than the develop the subject itself in detail, because the reaction proper-
ties are so sensitive to electronic structure, it may be appropriate in concluding
this section to illustrate the connection with a few examples. Effects arising
from differences in electronic structure are manifested in several different ways:
by affecting the rates of substitution, they can affect the choice of mechanism,
and, for an inner sphere reaction path, can determine whether binuclear
intermediates are easily observable, and whether there is net transfer of a group
from one center to another; even after the precursor complex is assembled,
orbital symmetry can affect the mechanism itself, as in the example offered in
an earlier paragraph, and can profoundly affect the rate of conversion of the
precursor to the successor complex.
The Cr(III)/Cr(II)
and
couples
offer perhaps the greatest contrasts in behavior. It should be noted that the